文章目录
- 红黑树
- 红黑树的概念
- 红黑树的性质
- 红黑树节点的定义
- 红黑树的插入
- 代码实现
- 总结
红黑树
AVL树是一颗绝对平衡的二叉搜索树,要求每个节点的左右高度差的绝对值不超过1,这样保证查询时的高效时间复杂度O( l o g 2 N ) log_2 N) log2N),但是要维护其绝对平衡,旋转的次数比较多。因此,如果一颗树的结构经常修改,那么AVL树就不太合适,所以就有了红黑树。
红黑树的概念
红黑树,是一种二叉搜索树,但在每个结点上增加一个存储位表示结点的颜色,可以是Red或Black。 通过对任何一条从根到叶子的路径上各个结点着色方式的限制,红黑树确保没有一条路径会比其他路径长出两倍,因而是接近平衡的。
红黑树的性质
- 每个节点不是红色就是黑色
- 根节点是黑色的
- 不存在连续的红色节点
- 任意一条从根到叶子的路径上的黑色节点的数量相同
根据上面的性质,红黑树就可以确保没有一条路径会比其他路径长出两倍,因为每条路径上的黑色节点的数量相同,所以理论上最短边一定都是黑色节点,最长边一定是一黑一红的不断重复的路径。
红黑树节点的定义
enum Color{RED,BLACK};template<class K, class V>struct RBTreeNode{RBTreeNode<K, V>* _left;RBTreeNode<K, V>* _right;RBTreeNode<K, V>* _parent;Color _col;pair<K, V> _kv;RBTreeNode(const pair<K, V>& kv):_left(nullptr),_right(nullptr),_parent(nullptr),_col(RED),_kv(kv){}};
插入新节点的颜色一定是红色,因为如果新节点的颜色是黑色,那么每条路径上的黑色节点的数量就不相同了,处理起来就比较麻烦,所以宁愿出现连续的红色节点,也不能让某一条路径上多出一个黑色节点。
红黑树的插入
1.根据二叉搜索树的规则插入新节点
bool Insert(const pair<K, V>& kv)
{if (_root == nullptr){_root = new Node(kv);_root->_col = BLACK;return true;}Node* curr = _root;Node* parent = nullptr;while (curr){if (curr->_kv.first < kv.first){parent = curr;curr = curr->_right;}else if (curr->_kv.first > kv.first){parent = curr;curr = curr->_left;}else{return false;}}curr = new Node(kv);if (parent->_kv.first < kv.first)parent->_right = curr;elseparent->_left = curr;curr->_parent = parent;
........
2.测新节点插入后,红黑树的性质是否造到破坏
bool Insert(const pair<K, V>& kv)
{if (_root == nullptr){_root = new Node(kv);_root->_col = BLACK;return true;}Node* curr = _root;Node* parent = nullptr;while (curr){if (curr->_kv.first < kv.first){parent = curr;curr = curr->_right;}else if (curr->_kv.first > kv.first){parent = curr;curr = curr->_left;}else{return false;}}curr = new Node(kv);if (parent->_kv.first < kv.first)parent->_right = curr;elseparent->_left = curr;curr->_parent = parent;while (parent && parent->_col == RED){Node* grandfather = parent->_parent;if (parent == grandfather->_left){Node* uncle = grandfather->_right;if (uncle && uncle->_col == RED){parent->_col = uncle->_col = BLACK;grandfather->_col = RED;curr = grandfather;parent = curr->_parent;}else{if (curr == parent->_left){// g// p u//cRotatoR(grandfather);parent->_col = BLACK;grandfather->_col = RED;}else{// g// p u// cRotatoL(parent);RotatoR(grandfather);curr->_col = BLACK;grandfather->_col = RED;}break;}}else{Node* uncle = grandfather->_left;if (uncle && uncle->_col == RED){parent->_col = uncle->_col = BLACK;grandfather->_col = RED;curr = grandfather;parent = curr->_parent;}else{if (curr == parent->_right){// g // u p// cRotatoL(grandfather);parent->_col = BLACK;grandfather->_col = RED;}else{// g // u p// cRotatoR(parent);RotatoL(grandfather);curr->_col = BLACK;grandfather->_col = RED;}break;}}}_root->_col = BLACK;return true;
}
void RotatoL(Node* parent)
{Node* subR = parent->_right;Node* subRL = subR->_left;parent->_right = subRL;if (subRL)subRL->_parent = parent;subR->_left = parent;Node* ppnode = parent->_parent;parent->_parent = subR;if (parent == _root){_root = subR;subR->_parent = nullptr;}else{if (ppnode->_left == parent)ppnode->_left = subR;elseppnode->_right = subR;subR->_parent = ppnode;}
}
void RotatoR(Node* parent)
{Node* subL = parent->_left;Node* subLR = subL->_right;parent->_left = subLR;if (subLR)subLR->_parent = parent;subL->_right = parent;Node* ppnode = parent->_parent;parent->_parent = subL;if (parent == _root){_root = subL;subL->_parent = nullptr;}else{if (ppnode->_left == parent)ppnode->_left = subL;elseppnode->_right = subL;subL->_parent = ppnode;}
}
代码实现
#pragma once
#include <utility>namespace lw
{enum Color{RED,BLACK};template<class K, class V>struct RBTreeNode{RBTreeNode<K, V>* _left;RBTreeNode<K, V>* _right;RBTreeNode<K, V>* _parent;Color _col;pair<K, V> _kv;RBTreeNode(const pair<K, V>& kv):_left(nullptr),_right(nullptr),_parent(nullptr),_col(RED),_kv(kv){}};template<class K, class V>class RBTree{typedef RBTreeNode<K, V> Node;public:bool Insert(const pair<K, V>& kv){if (_root == nullptr){_root = new Node(kv);_root->_col = BLACK;return true;}Node* curr = _root;Node* parent = nullptr;while (curr){if (curr->_kv.first < kv.first){parent = curr;curr = curr->_right;}else if (curr->_kv.first > kv.first){parent = curr;curr = curr->_left;}else{return false;}}curr = new Node(kv);if (parent->_kv.first < kv.first)parent->_right = curr;elseparent->_left = curr;curr->_parent = parent;while (parent && parent->_col == RED){Node* grandfather = parent->_parent;if (parent == grandfather->_left){Node* uncle = grandfather->_right;if (uncle && uncle->_col == RED){parent->_col = uncle->_col = BLACK;grandfather->_col = RED;curr = grandfather;parent = curr->_parent;}else{if (curr == parent->_left){// g// p u//cRotatoR(grandfather);parent->_col = BLACK;grandfather->_col = RED;}else{// g// p u// cRotatoL(parent);RotatoR(grandfather);curr->_col = BLACK;grandfather->_col = RED;}break;}}else{Node* uncle = grandfather->_left;if (uncle && uncle->_col == RED){parent->_col = uncle->_col = BLACK;grandfather->_col = RED;curr = grandfather;parent = curr->_parent;}else{if (curr == parent->_right){// g // u p// cRotatoL(grandfather);parent->_col = BLACK;grandfather->_col = RED;}else{// g // u p// cRotatoR(parent);RotatoL(grandfather);curr->_col = BLACK;grandfather->_col = RED;}break;}}}_root->_col = BLACK;return true;}void RotatoL(Node* parent){Node* subR = parent->_right;Node* subRL = subR->_left;parent->_right = subRL;if (subRL)subRL->_parent = parent;subR->_left = parent;Node* ppnode = parent->_parent;parent->_parent = subR;if (parent == _root){_root = subR;subR->_parent = nullptr;}else{if (ppnode->_left == parent)ppnode->_left = subR;elseppnode->_right = subR;subR->_parent = ppnode;}}void RotatoR(Node* parent){Node* subL = parent->_left;Node* subLR = subL->_right;parent->_left = subLR;if (subLR)subLR->_parent = parent;subL->_right = parent;Node* ppnode = parent->_parent;parent->_parent = subL;if (parent == _root){_root = subL;subL->_parent = nullptr;}else{if (ppnode->_left == parent)ppnode->_left = subL;elseppnode->_right = subL;subL->_parent = ppnode;}}void InOrder(){_InOrder(_root);}bool IsBalance(){if (_root && _root->_col == RED)return false;Node* left = _root;int count = 0;while (left){if (left->_col == BLACK)count++;left = left->_left;}return check(_root, 0, count);}private:bool check(Node* root, int count, int refBlackNumber){if (root == nullptr){if (count == refBlackNumber)return true;elsereturn false;}if (root->_col == RED && root->_parent->_col == RED)return false;if (root->_col == BLACK)count++;return check(root->_left, count, refBlackNumber)&& check(root->_right, count, refBlackNumber);}void _InOrder(Node* root){if (root == nullptr)return;_InOrder(root->_left);cout << root->_kv.first << " : " << root->_kv.second << endl;_InOrder(root->_right);}Node* _root = nullptr;};
}
总结
红黑树和AVL树都是高效的平衡二叉树,增删改查的时间复杂度都是O( l o g 2 N log_2 N log2N),红黑树不追求绝对平衡,其只需保证最长路径不超过最短路径的2倍,相对而言,降低了插入和旋转的次数,所以在经常进行增删的结构中性能比AVL树更优,而且红黑树实现比较简单,所以实际运用中红黑树更多。